Gene Regulatory Networks for Cell Reprogramming - A Boolean Networks Approach

Abstract: Maintenance of stem cell pluripotency and the differentiation of stem cells into specialised cells has been studied extensively. More recently, the possibility for a specialised cell to be transferred into a state of induced pluripotency has been demonstrated and investigated. Critical components of the gene regulatory networks governing these processes have been identified, but a complete model has yet to be developed. This thesis looks into the modelling of such networks by means of a deterministic Boolean networks approach. We addressed an experimental study on cell reprogramming of mouse embryonic fibroblast cells into induced pluripotent stem cells. In that study, barriers to reprogramming were identified as intermediate states between fibroblast and stem cell. We built a network topology on a subset of the genes considered in the study, inferring gene interactions from a public gene database, and trained an ensemble of networks based on available gene expression data. From this ensemble, relevant networks could be extracted that captured the observed barriers as attractors, along with the differentiated state and the stem cell state. Stability considerations indicated a linear ordering of these attractors consistent with the experimental findings. The networks were also able to capture reasonable reprogramming behaviour, for example transitions to the stem cell state upon overexpression of certain genes. In this context, a software environment was developed to study Boolean networks and their dynamics, with emphasis on gene regulatory networks. Features include tools for: extracting preliminary network topologies based on public gene databases, training Boolean networks based on such topologies and experimental gene expression data, analysing the stability of the networks, computing network attractors and searching for networks with certain attractor properties.